Use git attributes to determine generated and vendored status for language stats and diffs (#16773)

Replaces #16262
Replaces #16250
Replaces #14833

This PR first implements a `git check-attr` pipe reader - using `git check-attr --stdin -z --cached` - taking account of the change in the output format in git 1.8.5 and creates a helper function to read a tree into a temporary index file for that pipe reader.

It then wires this in to the language stats helper and into the git diff generation.

Files which are marked generated will be folded by default.

Fixes #14786
Fixes #12653

Author: zeripath

modules/analyze/generated.go

@@ -0,0 +1,28 @@
+// Copyright 2021 The Gitea Authors. All rights reserved.
+// Use of this source code is governed by a MIT-style
+// license that can be found in the LICENSE file.
+
+package analyze
+
+import (
+	"path/filepath"
+	"strings"
+
+	"github.com/go-enry/go-enry/v2/data"
+)
+
+// IsGenerated returns whether or not path is a generated path.
+func IsGenerated(path string) bool {
+	ext := strings.ToLower(filepath.Ext(path))
+	if _, ok := data.GeneratedCodeExtensions[ext]; ok {
+		return true
+	}
+
+	for _, m := range data.GeneratedCodeNameMatchers {
+		if m(path) {
+			return true
+		}
+	}
+
+	return false
+}

modules/git/repo_attribute.go

@@ -6,7 +6,12 @@ package git
 
 import (
 	"bytes"
+	"context"
 	"fmt"
+	"io"
+	"os"
+	"strconv"
+	"strings"
 )
 
 // CheckAttributeOpts represents the possible options to CheckAttribute
@@ -21,7 +26,7 @@ type CheckAttributeOpts struct {
 func (repo *Repository) CheckAttribute(opts CheckAttributeOpts) (map[string]map[string]string, error) {
 	err := LoadGitVersion()
 	if err != nil {
-		return nil, fmt.Errorf("Git version missing: %v", err)
+		return nil, fmt.Errorf("git version missing: %v", err)
 	}
 
 	stdOut := new(bytes.Buffer)
@@ -55,13 +60,14 @@ func (repo *Repository) CheckAttribute(opts CheckAttributeOpts) (map[string]map[
 	cmd := NewCommand(cmdArgs...)
 
 	if err := cmd.RunInDirPipeline(repo.Path, stdOut, stdErr); err != nil {
-		return nil, fmt.Errorf("Failed to run check-attr: %v\n%s\n%s", err, stdOut.String(), stdErr.String())
+		return nil, fmt.Errorf("failed to run check-attr: %v\n%s\n%s", err, stdOut.String(), stdErr.String())
 	}
 
+	// FIXME: This is incorrect on versions < 1.8.5
 	fields := bytes.Split(stdOut.Bytes(), []byte{'\000'})
 
 	if len(fields)%3 != 1 {
-		return nil, fmt.Errorf("Wrong number of fields in return from check-attr")
+		return nil, fmt.Errorf("wrong number of fields in return from check-attr")
 	}
 
 	var name2attribute2info = make(map[string]map[string]string)
@@ -80,3 +86,276 @@ func (repo *Repository) CheckAttribute(opts CheckAttributeOpts) (map[string]map[
 
 	return name2attribute2info, nil
 }
+
+// CheckAttributeReader provides a reader for check-attribute content that can be long running
+type CheckAttributeReader struct {
+	// params
+	Attributes []string
+	Repo       *Repository
+	IndexFile  string
+	WorkTree   string
+
+	stdinReader io.ReadCloser
+	stdinWriter *os.File
+	stdOut      attributeWriter
+	cmd         *Command
+	env         []string
+	ctx         context.Context
+	cancel      context.CancelFunc
+	running     chan struct{}
+}
+
+// Init initializes the cmd
+func (c *CheckAttributeReader) Init(ctx context.Context) error {
+	c.running = make(chan struct{})
+	cmdArgs := []string{"check-attr", "--stdin", "-z"}
+
+	if len(c.IndexFile) > 0 && CheckGitVersionAtLeast("1.7.8") == nil {
+		cmdArgs = append(cmdArgs, "--cached")
+		c.env = []string{"GIT_INDEX_FILE=" + c.IndexFile}
+	}
+
+	if len(c.WorkTree) > 0 && CheckGitVersionAtLeast("1.7.8") == nil {
+		c.env = []string{"GIT_WORK_TREE=" + c.WorkTree}
+	}
+
+	if len(c.Attributes) > 0 {
+		cmdArgs = append(cmdArgs, c.Attributes...)
+		cmdArgs = append(cmdArgs, "--")
+	} else {
+		lw := new(nulSeparatedAttributeWriter)
+		lw.attributes = make(chan attributeTriple)
+
+		c.stdOut = lw
+		c.stdOut.Close()
+		return fmt.Errorf("no provided Attributes to check")
+	}
+
+	c.ctx, c.cancel = context.WithCancel(ctx)
+	c.cmd = NewCommandContext(c.ctx, cmdArgs...)
+	var err error
+	c.stdinReader, c.stdinWriter, err = os.Pipe()
+	if err != nil {
+		return err
+	}
+
+	if CheckGitVersionAtLeast("1.8.5") == nil {
+		lw := new(nulSeparatedAttributeWriter)
+		lw.attributes = make(chan attributeTriple, 5)
+
+		c.stdOut = lw
+	} else {
+		lw := new(lineSeparatedAttributeWriter)
+		lw.attributes = make(chan attributeTriple, 5)
+
+		c.stdOut = lw
+	}
+	return nil
+}
+
+// Run run cmd
+func (c *CheckAttributeReader) Run() error {
+	stdErr := new(bytes.Buffer)
+	err := c.cmd.RunInDirTimeoutEnvFullPipelineFunc(c.env, -1, c.Repo.Path, c.stdOut, stdErr, c.stdinReader, func(_ context.Context, _ context.CancelFunc) error {
+		close(c.running)
+		return nil
+	})
+	defer c.cancel()
+	_ = c.stdOut.Close()
+	if err != nil && c.ctx.Err() != nil && err.Error() != "signal: killed" {
+		return fmt.Errorf("failed to run attr-check. Error: %w\nStderr: %s", err, stdErr.String())
+	}
+
+	return nil
+}
+
+// CheckPath check attr for given path
+func (c *CheckAttributeReader) CheckPath(path string) (map[string]string, error) {
+	select {
+	case <-c.ctx.Done():
+		return nil, c.ctx.Err()
+	case <-c.running:
+	}
+
+	if _, err := c.stdinWriter.Write([]byte(path + "\x00")); err != nil {
+		defer c.cancel()
+		return nil, err
+	}
+
+	if err := c.stdinWriter.Sync(); err != nil {
+		defer c.cancel()
+		return nil, err
+	}
+
+	rs := make(map[string]string)
+	for range c.Attributes {
+		select {
+		case attr := <-c.stdOut.ReadAttribute():
+			rs[attr.Attribute] = attr.Value
+		case <-c.ctx.Done():
+			return nil, c.ctx.Err()
+		}
+	}
+	return rs, nil
+}
+
+// Close close pip after use
+func (c *CheckAttributeReader) Close() error {
+	select {
+	case <-c.running:
+	default:
+		close(c.running)
+	}
+	defer c.cancel()
+	return c.stdinWriter.Close()
+}
+
+type attributeWriter interface {
+	io.WriteCloser
+	ReadAttribute() <-chan attributeTriple
+}
+
+type attributeTriple struct {
+	Filename  string
+	Attribute string
+	Value     string
+}
+
+type nulSeparatedAttributeWriter struct {
+	tmp        []byte
+	attributes chan attributeTriple
+	working    attributeTriple
+	pos        int
+}
+
+func (wr *nulSeparatedAttributeWriter) Write(p []byte) (n int, err error) {
+	l, read := len(p), 0
+
+	nulIdx := bytes.IndexByte(p, '\x00')
+	for nulIdx >= 0 {
+		wr.tmp = append(wr.tmp, p[:nulIdx]...)
+		switch wr.pos {
+		case 0:
+			wr.working = attributeTriple{
+				Filename: string(wr.tmp),
+			}
+		case 1:
+			wr.working.Attribute = string(wr.tmp)
+		case 2:
+			wr.working.Value = string(wr.tmp)
+		}
+		wr.tmp = wr.tmp[:0]
+		wr.pos++
+		if wr.pos > 2 {
+			wr.attributes <- wr.working
+			wr.pos = 0
+		}
+		read += nulIdx + 1
+		if l > read {
+			p = p[nulIdx+1:]
+			nulIdx = bytes.IndexByte(p, '\x00')
+		} else {
+			return l, nil
+		}
+	}
+	wr.tmp = append(wr.tmp, p...)
+	return len(p), nil
+}
+
+func (wr *nulSeparatedAttributeWriter) ReadAttribute() <-chan attributeTriple {
+	return wr.attributes
+}
+
+func (wr *nulSeparatedAttributeWriter) Close() error {
+	close(wr.attributes)
+	return nil
+}
+
+type lineSeparatedAttributeWriter struct {
+	tmp        []byte
+	attributes chan attributeTriple
+}
+
+func (wr *lineSeparatedAttributeWriter) Write(p []byte) (n int, err error) {
+	l := len(p)
+
+	nlIdx := bytes.IndexByte(p, '\n')
+	for nlIdx >= 0 {
+		wr.tmp = append(wr.tmp, p[:nlIdx]...)
+
+		if len(wr.tmp) == 0 {
+			// This should not happen
+			if len(p) > nlIdx+1 {
+				wr.tmp = wr.tmp[:0]
+				p = p[nlIdx+1:]
+				nlIdx = bytes.IndexByte(p, '\n')
+				continue
+			} else {
+				return l, nil
+			}
+		}
+
+		working := attributeTriple{}
+		if wr.tmp[0] == '"' {
+			sb := new(strings.Builder)
+			remaining := string(wr.tmp[1:])
+			for len(remaining) > 0 {
+				rn, _, tail, err := strconv.UnquoteChar(remaining, '"')
+				if err != nil {
+					if len(remaining) > 2 && remaining[0] == '"' && remaining[1] == ':' && remaining[2] == ' ' {
+						working.Filename = sb.String()
+						wr.tmp = []byte(remaining[3:])
+						break
+					}
+					return l, fmt.Errorf("unexpected tail %s", string(remaining))
+				}
+				_, _ = sb.WriteRune(rn)
+				remaining = tail
+			}
+		} else {
+			idx := bytes.IndexByte(wr.tmp, ':')
+			if idx < 0 {
+				return l, fmt.Errorf("unexpected input %s", string(wr.tmp))
+			}
+			working.Filename = string(wr.tmp[:idx])
+			if len(wr.tmp) < idx+2 {
+				return l, fmt.Errorf("unexpected input %s", string(wr.tmp))
+			}
+			wr.tmp = wr.tmp[idx+2:]
+		}
+
+		idx := bytes.IndexByte(wr.tmp, ':')
+		if idx < 0 {
+			return l, fmt.Errorf("unexpected input %s", string(wr.tmp))
+		}
+
+		working.Attribute = string(wr.tmp[:idx])
+		if len(wr.tmp) < idx+2 {
+			return l, fmt.Errorf("unexpected input %s", string(wr.tmp))
+		}
+
+		working.Value = string(wr.tmp[idx+2:])
+
+		wr.attributes <- working
+		wr.tmp = wr.tmp[:0]
+		if len(p) > nlIdx+1 {
+			p = p[nlIdx+1:]
+			nlIdx = bytes.IndexByte(p, '\n')
+			continue
+		} else {
+			return l, nil
+		}
+	}
+
+	wr.tmp = append(wr.tmp, p...)
+	return l, nil
+}
+
+func (wr *lineSeparatedAttributeWriter) ReadAttribute() <-chan attributeTriple {
+	return wr.attributes
+}
+
+func (wr *lineSeparatedAttributeWriter) Close() error {
+	close(wr.attributes)
+	return nil
+}